summaryrefslogtreecommitdiffstats
path: root/src/compiler
Commit message (Collapse)AuthorAgeFilesLines
* nir/algebraic: Make algebraic_parser_test.sh executable.Vinson Lee2018-12-061-0/+0
| | | | | | | | | | Fixes make check permission error. ../../bin/test-driver: line 107: ./nir/tests/algebraic_parser_test.sh: Permission denied FAIL nir/tests/algebraic_parser_test.sh (exit status: 126) Fixes: a0ae12ca91a4 ("nir/algebraic: Add unit tests for bitsize validation") Signed-off-by: Vinson Lee <[email protected]>
* nir: Make boolean conversions sized just like the othersJason Ekstrand2018-12-0513-45/+79
| | | | | | | | | Instead of a single i2b and b2i, we now have i2b32 and b2iN where N is one if 8, 16, 32, or 64. This leads to having a few more opcodes but now everything is consistent and booleans aren't a weird special case anymore. Reviewed-by: Connor Abbott <[email protected]>
* nir/opt_algebraic: Add 32-bit specifiers to a bunch of booleansJason Ekstrand2018-12-051-56/+56
| | | | Reviewed-by: Connor Abbott <[email protected]>
* nir/opt_algebraic: Drop bit-size suffixes from conversionsJason Ekstrand2018-12-051-29/+29
| | | | | | | | Suffixes are dropped from a bunch of conversion opcodes when it makes sense to do so. Others are kept if we really do want the bit-size restriction. Reviewed-by: Connor Abbott <[email protected]>
* nir/opt_algebraic: Simplify an optimization using the new search opsJason Ekstrand2018-12-051-7/+2
| | | | Reviewed-by: Connor Abbott <[email protected]>
* nir/algebraic: Add support for unsized conversion opcodesJason Ekstrand2018-12-053-10/+140
| | | | | | | | | | | | All conversion opcodes require a destination size but this makes constructing certain algebraic expressions rather cumbersome. This commit adds support to nir_search and nir_algebraic for writing conversion opcodes without a size. These meta-opcodes match any conversion of that type regardless of destination size and the size gets inferred from the sizes of the things being matched or from other opcodes in the expression. Reviewed-by: Connor Abbott <[email protected]>
* nir/algebraic: Refactor codegen a bitJason Ekstrand2018-12-051-10/+11
| | | | | | | Instead of using an OrderedDict, just have a (necessarily sorted) array of transforms and a set of opcodes. Reviewed-by: Connor Abbott <[email protected]>
* nir/algebraic: Clean up some __str__ cruftJason Ekstrand2018-12-051-4/+0
| | | | | | | Both of these things are already handled in the Value base class so we don't need to handle them explicitly in Constant. Reviewed-by: Connor Abbott <[email protected]>
* nir/opcodes: Rename tbool to tbool32Jason Ekstrand2018-12-051-11/+11
| | | | Reviewed-by: Connor Abbott <[email protected]>
* nir/opcodes: Pull in the type helpers from constant_expressionsJason Ekstrand2018-12-053-35/+35
| | | | | | | While we're at it, we rework them a bit to all use regular expressions and assert more. Reviewed-by: Connor Abbott <[email protected]>
* nir/algebraic: Add unit tests for bitsize validationConnor Abbott2018-12-054-1/+129
| | | | | | | | | The non-failure path can be tested by just compiling mesa and then testing it, but the failure paths won't be hit unless you make a mistake, so it's best to test them with some unit tests. Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Dylan Baker <[email protected]>
* nir/algebraic: Rewrite bit-size inferenceConnor Abbott2018-12-053-366/+317
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Before this commit, there were two copies of the algorithm: one in C, that we would use to figure out what bit-size to give the replacement expression, and one in Python, that emulated the C one and tried to prove that the C algorithm would never fail to correctly assign bit-sizes. That seemed pretty fragile, and likely to fall over if we make any changes. Furthermore, the C code was really just recomputing more-or-less the same thing as the Python code every time. Instead, we can just store the results of the Python algorithm in the C datastructure, and consult it to compute the bitsize of each value, moving the "brains" entirely into Python. Since the Python algorithm no longer has to match C, it's also a lot easier to change it to something more closely approximating an actual type-inference algorithm. The algorithm used is based on Hindley-Milner, although deliberately weakened a little. It's a few more lines than the old one, judging by the diffstat, but I think it's easier to verify that it's correct while being as general as possible. We could split this up into two changes, first making the C code use the results of the Python code and then rewriting the Python algorithm, but since the old algorithm never tracked which variable each equivalence class, it would mean we'd have to add some non-trivial code which would then get thrown away. I think it's better to see the final state all at once, although I could also try splitting it up. v2: - Replace instances of "== None" and "!= None" with "is None" and "is not None". - Rename first_src to first_unsized_src - Only merge the destination with the first unsized source, since the sources have already been merged. - Add a comment explaining what nir_search_value::bit_size now means. v3: - Fix one last instance to use "is not" instead of != - Don't try to be so clever when choosing which error message to print based on whether we're in the search or replace expression. - Fix trailing whitespace. Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Dylan Baker <[email protected]>
* spirv: add SpvCapabilityInt64AtomicsSamuel Pitoiset2018-12-052-1/+5
| | | | | | | Required for VK_KHR_shader_atomic_int64. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* mesa: expose AMD_texture_texture4Marek Olšák2018-12-043-0/+13
| | | | | | because the closed driver exposes it. Tested by piglit. Reviewed-by: Ilia Mirkin <[email protected]>
* mesa: Revert INTEL_fragment_shader_ordering supportMatt Turner2018-12-036-28/+0
| | | | | | | | | | | | | | | | This extension is not properly tested (testing for GL_ARB_fragment_shader_interlock is not sufficient), and since this was noted in review on August 28th no tests have been sent. Revert "i965: Add INTEL_fragment_shader_ordering support." Revert "mesa: Add GL/GLSL plumbing for INTEL_fragment_shader_ordering" This reverts commit 03ecec9ed2099f6e2b62994b33dc948dc731e7b8. This reverts commit 119435c8778dd26cb7c8bcde9f04b3982239fe60. Cc: [email protected] Acked-by: Jason Ekstrand <[email protected]> Acked-by: Eric Anholt <[email protected]>
* nir: Fix assert in print_intrinsic_instr().Józef Kucia2018-11-291-1/+1
| | | | | Signed-off-by: Józef Kucia <[email protected]> Reviewed-by: Lionel Landwerlin <[email protected]>
* nir/derefs: Add a nir_derefs_do_not_alias enum valueJason Ekstrand2018-11-282-3/+4
| | | | | | This makes some of the code more clear. Reviewed-by: Thomas Helland <[email protected]>
* glsl: Remove unused member variableMatt Turner2018-11-271-4/+0
| | | | Reviewed-by: Tapani Pälli <[email protected]>
* nir: Call fflush() at the end of nir_print_shader()Matt Turner2018-11-271-0/+1
| | | | | | | | We normally call with stderr which is unbuffered, so this won't affect that, but it does let me call nir_print_shader(nir, fopen("log", "w+")) from gdb and actually get the whole shader in my file. Reviewed-by: Tapani Pälli <[email protected]>
* nir: add fceil loweringJonathan Marek2018-11-272-0/+4
| | | | | | | | | lowers ceil(x) as -floor(-x) Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]> Signed-off-by: Rob Clark <[email protected]>
* glsl: free or reuse memory allocated for TF varyingGert Wollny2018-11-261-1/+8
| | | | | | | | | | | | | | | | | | | | | | | | When a shader program is de-serialized the gl_shader_program passed in may actually still hold memory allocations for the transform feedback varyings. If that is the case, free the varying names and reallocate the new storage for the names array. This fixes a memory leak: Direct leak of 48 byte(s) in 6 object(s) allocated from: in malloc (/usr/lib64/gcc/x86_64-pc-linux-gnu/7.3.0/libasan.so+0xdb880) in transform_feedback_varyings ../../samba/mesa/src/mesa/main/transformfeedback.c:875 in _mesa_TransformFeedbackVaryings ../../samba/mesa/src/mesa/main/transformfeedback.c:985 ... Indirect leak of 42 byte(s) in 6 object(s) allocated from: in __interceptor_strdup (/usr/lib64/gcc/x86_64-pc-linux-gnu/7.3.0/libasan.so+0x761c8) in transform_feedback_varyings ../../samba/mesa/src/mesa/main/transformfeedback.c:887 in _mesa_TransformFeedbackVaryings ../../samba/mesa/src/mesa/main/transformfeedback.c:985 Fixes: ab2643e4b06f63c93a57624003679903442634a8 glsl: serialize data from glTransformFeedbackVaryings Signed-off-by: Gert Wollny <[email protected]> Reviewed-by: Tapani Pälli <[email protected]>
* Revert "nir/builder: Assert that intN_t immediates fit"Iago Toral Quiroga2018-11-211-4/+0
| | | | | | | | | | | | | | | | | | | | | | | | This reverts commit 1f29f4db1e867357a119c0c7c34fb54dc27fb682. For this to work the compiler must ensure that it never puts the values that arrive to this helper into unsigned variables at any point in its processing, since that would not apply sign extension to the value and it would break the expectations here. Unfortunately, we use uint64_t extensively to pass and copy things around, so some times we get to this helper with values that are not properly sign extended to 64-bit. Here is an example for an 8-bit value that comes from a switch case: (gdb) p /x x $1 = 0xffffffd6 The value seems to have been sign extended to 32-bit at some point getting proper sign extension, but then copied into a uint64_t which wont' apply sign extension, breaking the expectations of the assertion. Reviewed-by: Juan A. Suarez <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir/from_ssa: fix bit-size of temporary registerIago Toral Quiroga2018-11-211-2/+5
| | | | Reviewed-by: Jason Ekstrand <[email protected]>
* meson: Add tests to suitesDylan Baker2018-11-203-7/+19
| | | | | | | | | | | | | | | | Meson test has a concepts of suites, which allow tests to be grouped together. This allows for a subtest of tests to be run only (say only the tests for nir). A test can be added to more than one suite, but for the most part I've only added a test to a single suite, though I've added a compiler group that includes nir, glsl, and glcpp tests. To use this you'll need to invoke meson test directly, instead of ninja test (which always runs all targets). it can be invoked as: `meson test -C builddir --suite $suitename` (meson test has addition options that are pretty useful). Tested-By: Gert Wollny <[email protected]> Acked-by: Eric Engestrom <[email protected]>
* nir: Make nir_lower_clip_vs optionally work with variables.Kenneth Graunke2018-11-192-12/+35
| | | | | | | | | | | | The way nir_lower_clip_vs() works with store_output intrinsics makes a ton of assumptions about the driver_location field. In i965 and iris, I'd rather do this lowering early and work with variables. v3d may want to switch to that as well, and ir3 could too, but I'm not sure exactly what would need updating. For now, handle both methods. Reviewed-by: Eric Anholt <[email protected]>
* nir: Save nir_variable pointers in nir_lower_clip_vs rather than locs.Kenneth Graunke2018-11-191-16/+8
| | | | | | I'll want the variables in the next patch. Reviewed-by: Eric Anholt <[email protected]>
* nir: Inline lower_clip_vs() into nir_lower_clip_vs().Kenneth Graunke2018-11-191-41/+32
| | | | | | It's now called exactly once, and there's not really any distinction. Reviewed-by: Eric Anholt <[email protected]>
* nir: Use nir_shader_get_entrypoint in nir_lower_clip_vs().Kenneth Graunke2018-11-191-4/+2
| | | | Reviewed-by: Eric Anholt <[email protected]>
* nir: handle shared pointers in lowering indirect derefs.Dave Airlie2018-11-201-2/+2
| | | | | | | Check if the base ends up with no variable, and continue if we see that case outside the loop. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: move getting deref from var after we check deref type.Dave Airlie2018-11-201-4/+3
| | | | | | | | | | I posted a load of hacks before to do this, Jason suggested this, just check the deref mode, not the variable mode and delay getting the variable until we know the type. avoids crashes when derefing shared memory pointers. Reviewed-by: Jason Ekstrand <[email protected]>
* spirv/vtn: handle variable pointers without offset loweringDave Airlie2018-11-202-8/+10
| | | | Reviewed-by: Jason Ekstrand <[email protected]>
* intel,nir: Move gl_LocalInvocationID lowering to nir_lower_system_valuesJason Ekstrand2018-11-192-1/+49
| | | | | | | | It's not at all intel-specific; the formula is dictated by OpenGL and Vulkan. The only intel-specific thing is that we need the lowering. As a nice side-effect, the new version is variable-group-size ready. Reviewed-by: Plamena Manolova <[email protected]>
* nir: Add alignment parameters to SSBO, UBO, and shared accessJason Ekstrand2018-11-157-10/+85
| | | | | | | | | | This also changes spirv_to_nir and glsl_to_nir to set them. The one place that doesn't set them is shared memory access lowering in nir_lower_io. That will have to be updated before any consumers of it can effectively use these new alignments. Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]> Acked-by: Karol Herbst <[email protected]>
* nir/lower_io: Add shared to get_io_offset_srcJason Ekstrand2018-11-151-0/+2
| | | | Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]>
* nir/glsl: Force 32-bit for UBO and SSBO BooleansJason Ekstrand2018-11-151-5/+12
| | | | Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]>
* nir/spirv: Force 32-bit for UBO and SSBO BooleansJason Ekstrand2018-11-151-2/+8
| | | | Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]>
* nir/builder: Add a nir_pack/unpack/bitcast helpersJason Ekstrand2018-11-152-76/+151
| | | | | | | | | The new helpers can generate any pack/unpack operation including those for which we do not have specific opcodes and they express a bitcast in terms of these pack/unpack operations. In particular, the new helpers properly handle 8-bit types. Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]>
* nir/builder: Add iadd_imm and imul_imm helpersJason Ekstrand2018-11-153-11/+20
| | | | | | | | | | The pattern of adding or multiplying an integer by an immediate is fairly common especially in deref chain handling. This adds a helper for it and uses it a few places. The advantage to the helper is that it automatically handles bit sizes for you. Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]> Reviewed-by: Karol Herbst <[email protected]>
* nir/builder: Assert that intN_t immediates fitJason Ekstrand2018-11-151-0/+4
| | | | | | | | This assert won't catch all mistakes with this helper but it will at least ensure that the top bits are all zero or all one which should help catch bugs. Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]>
* nir/lower_alu_to_scalar: Don't try to lower unpack_32_2x16Jason Ekstrand2018-11-151-0/+1
| | | | | | | It messes up when trying to lower. Cc: [email protected] Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]>
* glsl: Refactor type checking for redeclarationsIan Romanick2018-11-151-12/+7
| | | | | Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* glsl: Omit redundant qualifier checks on redeclarationsIan Romanick2018-11-151-6/+3
| | | | | Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* glsl: prevent qualifiers modification of predeclared variablesIan Romanick2018-11-151-24/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Section 3.7 (Identifiers) of the GLSL spec says: However, as noted in the specification, there are some cases where previously declared variables can be redeclared to change or add some property, and predeclared "gl_" names are allowed to be redeclared in a shader only for these specific purposes. More generally, it is an error to redeclare a variable, including those starting "gl_". This patch should fix piglit tests: clip-distance-redeclare-without-inout.frag clip-distance-redeclare-without-inout.vert However, this causes a regression in clip-distance-out-values.shader_test. A fix for that test has been sent to the piglit list for review: https://patchwork.freedesktop.org/patch/255201/ As far as I understood following mailing thread: https://lists.freedesktop.org/archives/piglit/2013-October/007935.html looks like we have accepted to remove an ability to change qualifiers but have not done it yet. Unless I missed something) v2 (idr): Move 'earlier->data.mode != var->data.mode' test much earlier in the function. Add special handling for gl_LastFragData. Signed-off-by: Andrii Simiklit <[email protected]> Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* compiler: avoid 'unused variable' warningsAndrii Simiklit2018-11-141-2/+1
| | | | | | | | | | | | | | | | | 1. nir/nir_lower_vars_to_ssa.c:691:21: warning: unused variable ‘var’ nir_variable *var = path->path[0]->var; v2: Changes for some part of 'may be used uninitialized' warnings were removed, seems like it is a compiler issue. ( Eric Engestrom <[email protected]> ) Possible like this one: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=46684 This issue is flagged as duplicate but an original one is not closed yet. Signed-off-by: Andrii Simiklit <[email protected]> Reviewed-by: Eric Engestrom <[email protected]>
* nir: Allow to skip integer ops in nir_lower_to_source_modsGert Wollny2018-11-142-31/+57
| | | | | | | | | | Some hardware supports source mods only for float operations. Make it possible to skip lowering to source mods in these cases. v2: use option flags instead of a boolean (Jason Ekstrand) Signed-off-by: Gert Wollny <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir/spirv: cast shift operand to u32Karol Herbst2018-11-142-0/+31
| | | | | | | | v2: fix for specialization constants as well Reviewed-by: Jason Ekstrand <[email protected]> Cc: [email protected] Signed-off-by: Karol Herbst <[email protected]>
* nir: replace nir_load_system_value calls with appropiate builder functionsKarol Herbst2018-11-142-4/+2
| | | | | | | | | this helps reduce the overall code changes when a bit_size parameter is added to nir_load_system_value Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Eric Anholt <[email protected]> Signed-off-by: Karol Herbst <[email protected]>
* nir: add const_index parameters to system value builder functionKarol Herbst2018-11-141-2/+19
| | | | | | | | | this allows to replace some nir_load_system_value calls with the specific system value constructor Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Eric Anholt <[email protected]> Signed-off-by: Karol Herbst <[email protected]>
* nir: Don't lower the local work group size if it's variable.Plamena Manolova2018-11-131-6/+18
| | | | | | | | | | If the local work group size is variable it won't be available at compile time so we can't lower it in nir_lower_system_values(). Signed-off-by: Plamena Manolova <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Karol Herbst <[email protected]>
* nir: add support for removing redundant stores to copy prop varTimothy Arceri2018-11-131-10/+35
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | For example the following type of thing is seen in TCS from a number of Vulkan and DXVK games: vec1 32 ssa_557 = deref_var &oPatch (shader_out float) vec1 32 ssa_558 = intrinsic load_deref (ssa_557) () vec1 32 ssa_559 = deref_var &oPatch@42 (shader_out float) vec1 32 ssa_560 = intrinsic load_deref (ssa_559) () vec1 32 ssa_561 = deref_var &oPatch@43 (shader_out float) vec1 32 ssa_562 = intrinsic load_deref (ssa_561) () intrinsic store_deref (ssa_557, ssa_558) (1) /* wrmask=x */ intrinsic store_deref (ssa_559, ssa_560) (1) /* wrmask=x */ intrinsic store_deref (ssa_561, ssa_562) (1) /* wrmask=x */ No shader-db changes on i965 (SKL). vkpipeline-db results RADV (VEGA): Totals from affected shaders: SGPRS: 7832 -> 7728 (-1.33 %) VGPRS: 6476 -> 6740 (4.08 %) Spilled SGPRs: 0 -> 0 (0.00 %) Spilled VGPRs: 0 -> 0 (0.00 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 0 -> 0 (0.00 %) dwords per thread Code Size: 469572 -> 456596 (-2.76 %) bytes LDS: 0 -> 0 (0.00 %) blocks Max Waves: 989 -> 960 (-2.93 %) Wait states: 0 -> 0 (0.00 %) The Max Waves and VGPRS changes here are misleading. What is happening is a bunch of TCS outputs are being optimised away as they are now recognised as unused. This results in more varyings being compacted via nir_compact_varyings() which can result in more register pressure when they are not packed in an optimal way. This is an existing problem independent of this patch. I've run some benchmarks and haven't noticed any performance regressions in affected games. Reviewed-by: Jason Ekstrand <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-19 23:14:42 +0000